diff options
Diffstat (limited to 'youtube_dl/extractor/youtube.py')
| -rw-r--r-- | youtube_dl/extractor/youtube.py | 319 |
1 files changed, 206 insertions, 113 deletions
diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index b252e36e1..9b39505ba 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -20,7 +20,6 @@ from ..compat import ( compat_urllib_parse_unquote, compat_urllib_parse_unquote_plus, compat_urllib_parse_urlparse, - compat_urllib_request, compat_urlparse, compat_str, ) @@ -35,6 +34,7 @@ from ..utils import ( orderedSet, parse_duration, remove_start, + sanitized_Request, smuggle_url, str_to_int, unescapeHTML, @@ -114,7 +114,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor): login_data = compat_urllib_parse.urlencode(encode_dict(login_form_strs)).encode('ascii') - req = compat_urllib_request.Request(self._LOGIN_URL, login_data) + req = sanitized_Request(self._LOGIN_URL, login_data) login_results = self._download_webpage( req, None, note='Logging in', errnote='unable to log in', fatal=False) @@ -147,7 +147,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor): tfa_data = compat_urllib_parse.urlencode(encode_dict(tfa_form_strs)).encode('ascii') - tfa_req = compat_urllib_request.Request(self._TWOFACTOR_URL, tfa_data) + tfa_req = sanitized_Request(self._TWOFACTOR_URL, tfa_data) tfa_results = self._download_webpage( tfa_req, None, note='Submitting TFA code', errnote='unable to submit tfa', fatal=False) @@ -178,6 +178,69 @@ class YoutubeBaseInfoExtractor(InfoExtractor): return +class YoutubeEntryListBaseInfoExtractor(InfoExtractor): + # Extract entries from page with "Load more" button + def _entries(self, page, playlist_id): + more_widget_html = content_html = page + for page_num in itertools.count(1): + for entry in self._process_page(content_html): + yield entry + + mobj = re.search(r'data-uix-load-more-href="/?(?P<more>[^"]+)"', more_widget_html) + if not mobj: + break + + more = self._download_json( + 'https://youtube.com/%s' % mobj.group('more'), playlist_id, + 'Downloading page #%s' % page_num, + transform_source=uppercase_escape) + content_html = more['content_html'] + if not content_html.strip(): + # Some webpages show a "Load more" button but they don't + # have more videos + break + more_widget_html = more['load_more_widget_html'] + + +class YoutubePlaylistBaseInfoExtractor(YoutubeEntryListBaseInfoExtractor): + def _process_page(self, content): + for video_id, video_title in self.extract_videos_from_page(content): + yield self.url_result(video_id, 'Youtube', video_id, video_title) + + def extract_videos_from_page(self, page): + ids_in_page = [] + titles_in_page = [] + for mobj in re.finditer(self._VIDEO_RE, page): + # The link with index 0 is not the first video of the playlist (not sure if still actual) + if 'index' in mobj.groupdict() and mobj.group('id') == '0': + continue + video_id = mobj.group('id') + video_title = unescapeHTML(mobj.group('title')) + if video_title: + video_title = video_title.strip() + try: + idx = ids_in_page.index(video_id) + if video_title and not titles_in_page[idx]: + titles_in_page[idx] = video_title + except ValueError: + ids_in_page.append(video_id) + titles_in_page.append(video_title) + return zip(ids_in_page, titles_in_page) + + +class YoutubePlaylistsBaseInfoExtractor(YoutubeEntryListBaseInfoExtractor): + def _process_page(self, content): + for playlist_id in re.findall(r'href="/?playlist\?list=(.+?)"', content): + yield self.url_result( + 'https://www.youtube.com/playlist?list=%s' % playlist_id, 'YoutubePlaylist') + + def _real_extract(self, url): + playlist_id = self._match_id(url) + webpage = self._download_webpage(url, playlist_id) + title = self._og_search_title(webpage, fatal=False) + return self.playlist_result(self._entries(webpage, playlist_id), playlist_id, title) + + class YoutubeIE(YoutubeBaseInfoExtractor): IE_DESC = 'YouTube.com' _VALID_URL = r"""(?x)^ @@ -195,7 +258,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): |(?: # or the v= param in all its forms (?:(?:watch|movie)(?:_popup)?(?:\.php)?/?)? # preceding watch(_popup|.php) or nothing (like /?v=xxxx) (?:\?|\#!?) # the params delimiter ? or # or #! - (?:.*?&)?? # any other preceding param (like /?s=tuff&v=xxxx) + (?:.*?[&;])?? # any other preceding param (like /?s=tuff&v=xxxx or ?s=tuff&v=V36LpHqtcDY) v= ) )) @@ -283,6 +346,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): '247': {'ext': 'webm', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, '248': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, '271': {'ext': 'webm', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, + # itag 272 videos are either 3840x2160 (e.g. RtoitU2A-3E) or 7680x4320 (sLprVF6d7Ug) '272': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, '302': {'ext': 'webm', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'vp9'}, '303': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'vp9'}, @@ -363,7 +427,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'title': 'Principal Sexually Assaults A Teacher - Episode 117 - 8th June 2012', 'description': 'md5:09b78bd971f1e3e289601dfba15ca4f7', 'uploader': 'SET India', - 'uploader_id': 'setindia' + 'uploader_id': 'setindia', + 'age_limit': 18, } }, { @@ -500,7 +565,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'info_dict': { 'id': 'lqQg6PlCWgI', 'ext': 'mp4', - 'upload_date': '20120724', + 'upload_date': '20150827', 'uploader_id': 'olympic', 'description': 'HO09 - Women - GER-AUS - Hockey - 31 July 2012 - London 2012 Olympic Games', 'uploader': 'Olympics', @@ -628,6 +693,47 @@ class YoutubeIE(YoutubeBaseInfoExtractor): { 'url': 'http://vid.plus/FlRa-iH7PGw', 'only_matching': True, + }, + { + # Title with JS-like syntax "};" (see https://github.com/rg3/youtube-dl/issues/7468) + 'url': 'https://www.youtube.com/watch?v=lsguqyKfVQg', + 'info_dict': { + 'id': 'lsguqyKfVQg', + 'ext': 'mp4', + 'title': '{dark walk}; Loki/AC/Dishonored; collab w/Elflover21', + 'description': 'md5:8085699c11dc3f597ce0410b0dcbb34a', + 'upload_date': '20151119', + 'uploader_id': 'IronSoulElf', + 'uploader': 'IronSoulElf', + }, + 'params': { + 'skip_download': True, + }, + }, + { + # Tags with '};' (see https://github.com/rg3/youtube-dl/issues/7468) + 'url': 'https://www.youtube.com/watch?v=Ms7iBXnlUO8', + 'only_matching': True, + }, + { + # Video with yt:stretch=17:0 + 'url': 'https://www.youtube.com/watch?v=Q39EVAstoRM', + 'info_dict': { + 'id': 'Q39EVAstoRM', + 'ext': 'mp4', + 'title': 'Clash Of Clans#14 Dicas De Ataque Para CV 4', + 'description': 'md5:ee18a25c350637c8faff806845bddee9', + 'upload_date': '20151107', + 'uploader_id': 'UCCr7TALkRbo3EtFzETQF1LA', + 'uploader': 'CH GAMER DROID', + }, + 'params': { + 'skip_download': True, + }, + }, + { + 'url': 'https://www.youtube.com/watch?feature=player_embedded&amp;v=V36LpHqtcDY', + 'only_matching': True, } ] @@ -657,7 +763,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): def _extract_signature_function(self, video_id, player_url, example_sig): id_m = re.match( - r'.*?-(?P<id>[a-zA-Z0-9_-]+)(?:/watch_as3|/html5player(?:-new)?)?\.(?P<ext>[a-z]+)$', + r'.*?-(?P<id>[a-zA-Z0-9_-]+)(?:/watch_as3|/html5player(?:-new)?|/base)?\.(?P<ext>[a-z]+)$', player_url) if not id_m: raise ExtractorError('Cannot identify player %r' % player_url) @@ -812,16 +918,33 @@ class YoutubeIE(YoutubeBaseInfoExtractor): return {} return sub_lang_list + def _get_ytplayer_config(self, video_id, webpage): + patterns = ( + # User data may contain arbitrary character sequences that may affect + # JSON extraction with regex, e.g. when '};' is contained the second + # regex won't capture the whole JSON. Yet working around by trying more + # concrete regex first keeping in mind proper quoted string handling + # to be implemented in future that will replace this workaround (see + # https://github.com/rg3/youtube-dl/issues/7468, + # https://github.com/rg3/youtube-dl/pull/7599) + r';ytplayer\.config\s*=\s*({.+?});ytplayer', + r';ytplayer\.config\s*=\s*({.+?});', + ) + config = self._search_regex( + patterns, webpage, 'ytplayer.config', default=None) + if config: + return self._parse_json( + uppercase_escape(config), video_id, fatal=False) + def _get_automatic_captions(self, video_id, webpage): """We need the webpage for getting the captions url, pass it as an argument to speed up the process.""" self.to_screen('%s: Looking for automatic captions' % video_id) - mobj = re.search(r';ytplayer.config = ({.*?});', webpage) + player_config = self._get_ytplayer_config(video_id, webpage) err_msg = 'Couldn\'t find automatic captions for %s' % video_id - if mobj is None: + if not player_config: self._downloader.report_warning(err_msg) return {} - player_config = json.loads(mobj.group(1)) try: args = player_config['args'] caption_url = args['ttsurl'] @@ -1028,10 +1151,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor): age_gate = False video_info = None # Try looking directly into the video webpage - mobj = re.search(r';ytplayer\.config\s*=\s*({.*?});', video_webpage) - if mobj: - json_code = uppercase_escape(mobj.group(1)) - ytplayer_config = json.loads(json_code) + ytplayer_config = self._get_ytplayer_config(video_id, video_webpage) + if ytplayer_config: args = ytplayer_config['args'] if args.get('url_encoded_fmt_stream_map'): # Convert to the same format returned by compat_parse_qs @@ -1061,6 +1182,17 @@ class YoutubeIE(YoutubeBaseInfoExtractor): if not video_info: video_info = get_video_info if 'token' in get_video_info: + # Different get_video_info requests may report different results, e.g. + # some may report video unavailability, but some may serve it without + # any complaint (see https://github.com/rg3/youtube-dl/issues/7362, + # the original webpage as well as el=info and el=embedded get_video_info + # requests report video unavailability due to geo restriction while + # el=detailpage succeeds and returns valid data). This is probably + # due to YouTube measures against IP ranges of hosting providers. + # Working around by preferring the first succeeded video_info containing + # the token if no such video_info yet was found. + if 'token' not in video_info: + video_info = get_video_info break if 'token' not in video_info: if 'reason' in video_info: @@ -1286,7 +1418,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): player_desc = 'flash player %s' % player_version else: player_version = self._search_regex( - r'html5player-([^/]+?)(?:/html5player(?:-new)?)?\.js', + [r'html5player-([^/]+?)(?:/html5player(?:-new)?)?\.js', r'(?:www|player)-([^/]+)/base\.js'], player_url, 'html5 player', fatal=False) player_desc = 'html5 player %s' % player_version @@ -1348,6 +1480,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor): manifest_url = video_info['hlsvp'][0] url_map = self._extract_from_m3u8(manifest_url, video_id) formats = _map_to_format_list(url_map) + # Accept-Encoding header causes failures in live streams on Youtube and Youtube Gaming + for a_format in formats: + a_format.setdefault('http_headers', {})['Youtubedl-no-compression'] = 'True' else: raise ExtractorError('no conn, hlsvp or url_encoded_fmt_stream_map information found in video info') @@ -1385,10 +1520,15 @@ class YoutubeIE(YoutubeBaseInfoExtractor): r'<meta\s+property="og:video:tag".*?content="yt:stretch=(?P<w>[0-9]+):(?P<h>[0-9]+)">', video_webpage) if stretched_m: - ratio = float(stretched_m.group('w')) / float(stretched_m.group('h')) - for f in formats: - if f.get('vcodec') != 'none': - f['stretched_ratio'] = ratio + w = float(stretched_m.group('w')) + h = float(stretched_m.group('h')) + # yt:stretch may hold invalid ratio data (e.g. for Q39EVAstoRM ratio is 17:0). + # We will only process correct ratios. + if w > 0 and h > 0: + ratio = w / h + for f in formats: + if f.get('vcodec') != 'none': + f['stretched_ratio'] = ratio self._sort_formats(formats) @@ -1419,7 +1559,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): } -class YoutubePlaylistIE(YoutubeBaseInfoExtractor): +class YoutubePlaylistIE(YoutubeBaseInfoExtractor, YoutubePlaylistBaseInfoExtractor): IE_DESC = 'YouTube.com playlists' _VALID_URL = r"""(?x)(?: (?:https?://)? @@ -1427,7 +1567,7 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor): youtube\.com/ (?: (?:course|view_play_list|my_playlists|artist|playlist|watch|embed/videoseries) - \? (?:.*?&)*? (?:p|a|list)= + \? (?:.*?[&;])*? (?:p|a|list)= | p/ ) ( @@ -1440,7 +1580,7 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor): ((?:PL|LL|EC|UU|FL|RD|UL)[0-9A-Za-z-_]{10,}) )""" _TEMPLATE_URL = 'https://www.youtube.com/playlist?list=%s' - _VIDEO_RE = r'href="\s*/watch\?v=(?P<id>[0-9A-Za-z_-]{11})&[^"]*?index=(?P<index>\d+)' + _VIDEO_RE = r'href="\s*/watch\?v=(?P<id>[0-9A-Za-z_-]{11})&[^"]*?index=(?P<index>\d+)(?:[^>]+>(?P<title>[^<]+))?' IE_NAME = 'youtube:playlist' _TESTS = [{ 'url': 'https://www.youtube.com/playlist?list=PLwiyx1dc3P2JR9N8gQaQN_BCvlSlap7re', @@ -1557,37 +1697,11 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor): else: self.report_warning('Youtube gives an alert message: ' + match) - # Extract the video ids from the playlist pages - def _entries(): - more_widget_html = content_html = page - for page_num in itertools.count(1): - matches = re.finditer(self._VIDEO_RE, content_html) - # We remove the duplicates and the link with index 0 - # (it's not the first video of the playlist) - new_ids = orderedSet(m.group('id') for m in matches if m.group('index') != '0') - for vid_id in new_ids: - yield self.url_result(vid_id, 'Youtube', video_id=vid_id) - - mobj = re.search(r'data-uix-load-more-href="/?(?P<more>[^"]+)"', more_widget_html) - if not mobj: - break - - more = self._download_json( - 'https://youtube.com/%s' % mobj.group('more'), playlist_id, - 'Downloading page #%s' % page_num, - transform_source=uppercase_escape) - content_html = more['content_html'] - if not content_html.strip(): - # Some webpages show a "Load more" button but they don't - # have more videos - break - more_widget_html = more['load_more_widget_html'] - playlist_title = self._html_search_regex( - r'(?s)<h1 class="pl-header-title[^"]*">\s*(.*?)\s*</h1>', + r'(?s)<h1 class="pl-header-title[^"]*"[^>]*>\s*(.*?)\s*</h1>', page, 'title') - return self.playlist_result(_entries(), playlist_id, playlist_title) + return self.playlist_result(self._entries(page, playlist_id), playlist_id, playlist_title) def _real_extract(self, url): # Extract playlist id @@ -1613,36 +1727,31 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor): return self._extract_playlist(playlist_id) -class YoutubeChannelIE(InfoExtractor): +class YoutubeChannelIE(YoutubePlaylistBaseInfoExtractor): IE_DESC = 'YouTube.com channels' _VALID_URL = r'https?://(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/(?P<id>[0-9A-Za-z_-]+)' _TEMPLATE_URL = 'https://www.youtube.com/channel/%s/videos' + _VIDEO_RE = r'(?:title="(?P<title>[^"]+)"[^>]+)?href="/watch\?v=(?P<id>[0-9A-Za-z_-]+)&?' IE_NAME = 'youtube:channel' _TESTS = [{ 'note': 'paginated channel', 'url': 'https://www.youtube.com/channel/UCKfVa3S1e4PHvxWcwyMMg8w', 'playlist_mincount': 91, 'info_dict': { - 'id': 'UCKfVa3S1e4PHvxWcwyMMg8w', + 'id': 'UUKfVa3S1e4PHvxWcwyMMg8w', + 'title': 'Uploads from lex will', } + }, { + 'note': 'Age restricted channel', + # from https://www.youtube.com/user/DeusExOfficial + 'url': 'https://www.youtube.com/channel/UCs0ifCMCm1icqRbqhUINa0w', + 'playlist_mincount': 64, + 'info_dict': { + 'id': 'UUs0ifCMCm1icqRbqhUINa0w', + 'title': 'Uploads from Deus Ex', + }, }] - @staticmethod - def extract_videos_from_page(page): - ids_in_page = [] - titles_in_page = [] - for mobj in re.finditer(r'(?:title="(?P<title>[^"]+)"[^>]+)?href="/watch\?v=(?P<id>[0-9A-Za-z_-]+)&?', page): - video_id = mobj.group('id') - video_title = unescapeHTML(mobj.group('title')) - try: - idx = ids_in_page.index(video_id) - if video_title and not titles_in_page[idx]: - titles_in_page[idx] = video_title - except ValueError: - ids_in_page.append(video_id) - titles_in_page.append(video_title) - return zip(ids_in_page, titles_in_page) - def _real_extract(self, url): channel_id = self._match_id(url) @@ -1661,7 +1770,7 @@ class YoutubeChannelIE(InfoExtractor): 'channelId', channel_page, 'channel id', default=None) if not channel_playlist_id: channel_playlist_id = self._search_regex( - r'data-channel-external-id="([^"]+)"', + r'data-(?:channel-external-|yt)id="([^"]+)"', channel_page, 'channel id', default=None) if channel_playlist_id and channel_playlist_id.startswith('UC'): playlist_id = 'UU' + channel_playlist_id[2:] @@ -1685,29 +1794,7 @@ class YoutubeChannelIE(InfoExtractor): for video_id, video_title in self.extract_videos_from_page(channel_page)] return self.playlist_result(entries, channel_id) - def _entries(): - more_widget_html = content_html = channel_page - for pagenum in itertools.count(1): - - for video_id, video_title in self.extract_videos_from_page(content_html): - yield self.url_result( - video_id, 'Youtube', video_id=video_id, - video_title=video_title) - - mobj = re.search( - r'data-uix-load-more-href="/?(?P<more>[^"]+)"', - more_widget_html) - if not mobj: - break - - more = self._download_json( - 'https://youtube.com/%s' % mobj.group('more'), channel_id, - 'Downloading page #%s' % (pagenum + 1), - transform_source=uppercase_escape) - content_html = more['content_html'] - more_widget_html = more['load_more_widget_html'] - - return self.playlist_result(_entries(), channel_id) + return self.playlist_result(self._entries(channel_page, channel_id), channel_id) class YoutubeUserIE(YoutubeChannelIE): @@ -1738,6 +1825,29 @@ class YoutubeUserIE(YoutubeChannelIE): return super(YoutubeUserIE, cls).suitable(url) +class YoutubeUserPlaylistsIE(YoutubePlaylistsBaseInfoExtractor): + IE_DESC = 'YouTube.com user playlists' + _VALID_URL = r'https?://(?:\w+\.)?youtube\.com/user/(?P<id>[^/]+)/playlists' + IE_NAME = 'youtube:user:playlists' + + _TESTS = [{ + 'url': 'http://www.youtube.com/user/ThirstForScience/playlists', + 'playlist_mincount': 4, + 'info_dict': { + 'id': 'ThirstForScience', + 'title': 'Thirst for Science', + }, + }, { + # with "Load more" button + 'url': 'http://www.youtube.com/user/igorkle1/playlists?view=1&sort=dd', + 'playlist_mincount': 70, + 'info_dict': { + 'id': 'igorkle1', + 'title': 'Игорь Клейнер', + }, + }] + + class YoutubeSearchIE(SearchInfoExtractor, YoutubePlaylistIE): IE_DESC = 'YouTube.com searches' # there doesn't appear to be a real limit, for example if you search for @@ -1833,7 +1943,7 @@ class YoutubeSearchURLIE(InfoExtractor): } -class YoutubeShowIE(InfoExtractor): +class YoutubeShowIE(YoutubePlaylistsBaseInfoExtractor): IE_DESC = 'YouTube.com (multi-season) shows' _VALID_URL = r'https?://www\.youtube\.com/show/(?P<id>[^?#]*)' IE_NAME = 'youtube:show' @@ -1847,26 +1957,9 @@ class YoutubeShowIE(InfoExtractor): }] def _real_extract(self, url): - mobj = re.match(self._VALID_URL, url) - playlist_id = mobj.group('id') - webpage = self._download_webpage( - 'https://www.youtube.com/show/%s/playlists' % playlist_id, playlist_id, 'Downloading show webpage') - # There's one playlist for each season of the show - m_seasons = list(re.finditer(r'href="(/playlist\?list=.*?)"', webpage)) - self.to_screen('%s: Found %s seasons' % (playlist_id, len(m_seasons))) - entries = [ - self.url_result( - 'https://www.youtube.com' + season.group(1), 'YoutubePlaylist') - for season in m_seasons - ] - title = self._og_search_title(webpage, fatal=False) - - return { - '_type': 'playlist', - 'id': playlist_id, - 'title': title, - 'entries': entries, - } + playlist_id = self._match_id(url) + return super(YoutubeShowIE, self)._real_extract( + 'https://www.youtube.com/show/%s/playlists' % playlist_id) class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor): |